The recently formed LIDA Environment theme has started fostering cross-university collaboration between researchers at Leeds University who apply machine learning in their day-to-day research in the physical sciences. This cross-discipline collaboration has crystallised around the recently-formed Scientific Machine Learning (SciML) Community which brings together in particular early-career machine learning practitioners at monthly informal ML meet-ups now hosted within LIDA (first Friday of every month at 11am, everyone welcome!).
At a full-day workshop hosted and facilitated by LIDA in July the SciML community collectively analysed how research is currently carried out and what is needed by current ML researchers to continue to produce ground-breaking research at Leeds. From this discussion a detailed plan was produced, which highlights the need for
1) GPU-enabled low-latency computing infrastruture for model development
2) training in state-of-the-art techniques being developed within the machine learning community.
Although the environment theme has only recently begun, we have a plan of some activities to promote engagement and targets for the coming year.
– To grow and promote the SciML group, by financially supporting PGRs to lead discussion sessions and organise seminars
– We aim to advertise several PGR bursaries for travel, to support LIDA environment-facing research alongside a short early-career researcher visitor scheme
– We will highlight the advances necessary in improving computational hardware to enable better research, to feed into the University’s infrastructure plans
– We aim to expand the set of training materials in scientific machine learning to support a growing community of users